Recently researchers have identified a group of children with language impairment (LI) whose IQ scores are below the typical IQ cutoff of 85 for specific language impairment (SLI) but above the IQ cutoff of 70 for intellectual disability (Weismer et al., 2000). This group is referred to as having nonspecific language impairment (NLI). Eleven children with NLI, eleven same-age peers with SLI, and eleven same-age typical peers were compared on withdrawn and sociable behaviors demonstrated in the classroom setting using the Teacher Behavior Rating Scale (Hart & Robinson, 1996). Three subtypes of withdrawal (solitary-active, solitary-passive, reticent) and two subtypes of sociable behavior (impulse control/likability and prosocial) were examined. The children were compared on the severity of their ratings as well as the pattern of item responses for each behavioral subtype. Teachers rated the group with NLI significantly more poorly than the typical group on the reticence, impulse control/likability, and prosocial subtypes. Teachers rated the group with SLI significantly more poorly than the typical group on the impulse control/likability subtype. Participants were added to the typical group and the group with SLI to increase statistical power. With additional participants, the children with SLI were also rated significantly more poorly than the typical children on the reticence, likability, and prosocial subtypes. The group with NLI was consistently rated more poorly than the group with SLI except on the solitary-active subtype. However, there were no significant differences between the groups with LI, even with additional participants. The pattern of item responses was similar between the groups with NLI and SLI on all behavioral subtypes. In addition, the pattern of item responses for the groups with LI also matched the pattern of the typical group on the solitary-active withdrawal, impulse control/likability, and prosocial subtypes. These data indicate that the differences between the groups with NLI and SLI may be quantitative but not qualitative.